Schizophrenia (SCZ) is a highly heritable psychiatric disorder with an elusive pathophysiology and few novel treatments under development. Our understanding of the biological basis of SCZ would be greatly accelerated by the isolation of a substantial number of precise causal mutations that influence risk. Copy number variants (CNVs) are an attractive source for causative mutations since, by altering gene dosage or structure, they provide a clear direction of effect and molecular mechanism. We propose to leverage the power of multiple technologies and large samples used in the Large-Scale Swedish Schizophrenia Association Study funded by NIMH to identify gene- and exon-level CNVs associated with SCZ. Our overarching goal is to discover risk loci resulting from changes in copy number in specific genes that could lead to improved understanding of SCZ. First, we will combine multiple CNV call sets from complimentary technologies to identify gene- and exon-level CNVs nominally associated with SCZ. Data available include genome-wide SNP, exome genotyping arrays and exome sequencing in 5,001 SCZ cases and 6,234 controls. Second, we will integrate the normalized raw data from all three datasets to construct a virtual array and use a specialized hidden Markov model (VAMM) to survey genic CNVs to yield high-density CNV calls. Third, we will create a prioritized list of gene/exon level CNV associations with SCZ and validate 50 putative CNVs using a precise and high sensitivity independent methodology. Finally, we will evaluate whether validated CNV regions associate with SCZ beyond chance by querying existing genomic data available from a distinct set of 20,000 cases and 20,000 controls. To date, only a few specific genes have been implicated via CNV analysis, all of which provided novel insights into SCZ pathophysiology. We contend that these genes are the tip of an iceberg and the current deficiency is simply attributed to the lack of methods for gene-focused CNV evaluation. If this R21 study can identify even one new exon- or gene-level CNV whose dosage alters SCZ risk, it will represent an important advance in our understanding of the biological basis of SCZ. Any identified CNV would provide the basis for an R01 application to understand the mechanisms linking its dosage to SCZ risk.